Clin. Exp. Metastasis, 1996, 14, 115-124

Effect of matrix metalloproteinase inhibitors on tumor growth and spontaneous metastasis James G. Conway, Suzanne J. Trexler, Jean A. Wakefield, Brian E. Marron*, David L. Emerson, David M. Bickett#, David N. Deaton*, Deanna Garrison*, Mike Elder*§, Andy McElroy*§, Neville Willmott:l:, Andrew J. P. Dockerty$ and Gerard M. McGeehant Departments of Pharmacology, *Medicinal Chemistry and tBiochemistry, Glaxo Inc., Research Triangle Park, NC, USA; and ~Celltech Therapeutics Ltd, Slough, UK (Received 7 July 1995; accepted in revised form 19 October 1995) Four potent, synthetic inhibitors of matrix metalloproteinases (MMPs) were assessed as inhibitors of tumor growth and spontaneous metastasis to the lung. Mat Ly Lu rat prostate tumor, LOX human melanoma and M27 murine Lewis lung tumor were implanted subcutaneously (s.c.) in mice and allowed to grow for 3-12 days. The lungs of the tumor-bearing mice were then removed and implanted s.c. into untreated mice, and the outgrowth of secondary tumors from the implanted lungs measured. The incidence and rate of outgrowth of secondary tumors increased with the length of primary tumor growth, validating these measurements as indices of spontaneous metastasis to the lung. Compounds were tested by s.c. implantation of minipumps which delivered compound throughout the period of primary tumor growth and spontaneous metastasis to the lung at steady-state drug concentrations orders of magnitude greater than the concentrations needed to either inhibit collagenase, gelatinase or stromelysin in vitro. Inhibitor treatment slowed the growth of primary s.c. Mat Ly Lu and LOX tumors by 40-60% but had no significant effect on the growth of primary M27 tumors. Surprisingly, inhibitor treatment had no significant effect on the ability of the lung to generate secondary tumors when reimplanted s.c. in untreated mice. Because of the possible importance of cathepsins B, H and L in tumor growth and metastasis, the irreversible inhibitor E-64 was also infused by s.c. minipump. E-64 had no effect on the growth or spontanous metastasis of Mat Ly Lu or M27 tumors. Keywords: collagenases, E64, GI168, GI 173, GI 179, GI 184, gelatinase, metalloproteinase inhibitors, metastasis, neutral metalloproteinases, stromelysin

Introduction Several steps in tumor growth and metastasis require proteolytic degradation of extracellular matrix in the basement membrane of vessels and in the interstitial space. Rapid tumor growth requires neovascularization [1]. This involves degradation of the basement membrane of vessels as endothelial cells start to migrate toward the tumor to form new vessels. The capacity of some tumors to invade local tissues and penetrate blood vessels and lymphatics also involves Address correspondence to: J. G. Conway, Glaxo Inc., 5 Moore Drive, Research Triangle Park, NC 27709, USA. Tel: (+ 1) 919 941 3508; Fax: (+1) 919 941 3777. ~Current addresses: M. Elder, Himont USA Inc., 12001 Bay Area Blvd, Pasedena, TX 77507, USA; A. McElroy, Discovery Research, Pfizer Central Research, Sandwich CT13 9NJ, UK.

© 1996 Rapid Science Publishers

proteolytic degradation of extracellular matrix. Once dispersed by blood and lymph to distant sites, tumor cells use proteolytic enzymes to penetrate the vessel basement membrane and migrate into the tissue to form secondary tumors [2]. The enzymes most implicated in matrix degradation include matrix metalloproteinases (MMPs), plasminogen activators and cathepsins [3,4]. MMPs are a family of zinc-dependent endopeptidases that degrade various components of the extracellular matrix [3]: collagenase (MMP-1) degrades collagen type I; gelatinase (MMP-2 and 9) degrades collagen types IV and V; and stromelysin (MMP-3) degrades proteoglycans and fibronectin as well as collagen type IV. In normal mature tissues, MMPs are expressed at low or undetectable levels. However, in situations such as tumor angiogenesis, tumor invasion and metastasis, Clinical & Experimental Metastasis Vol 14 No 2 115

C. Clemente et al.

individual cells or foci that are actively penetrating surrounding tissue or vessels, and thus degrading matrix barriers, often show increased expression of MMPs in histological preparations [5-10]. The gelatinases and stromelysin-3 are all reported to be negative prognostic indicators in clinicial samples of human tumors [5-10]. Furthermore, in experimental tumor systems, correlations have been observed between the capacity of cells to secrete MMPs in vitro and to metastasize in vivo [11-15]. Tissue inhibitor of metaUoproteinases (TIMP) and synthetic inhibitors have been used recently to evaluate the importance of MMPs as therapeutic targets in models of tumor growth and metastasis. Recombinant TIM P partially decreased the ability of B 16/F 10 tumor cells to colonize mouse lungs after intravenous (i.v.) injection (experimental metastasis) [16, 17], but had no effect on the growth of subcutaneous (s.c.) tumors [17]. Expression of TIMP-1 in B16/F10 cells decreased the growth of the cells in vitro and inhibited experimental metastasis [18]. TIMP-1 expression also decreased tumor take after s.c. injection, but had little effect on the growth of the tumors that did establish [18]. High TIMP-2 expression inhibited the experimental metastasis of c-Ha-ras rat embryo cells, whereas both moderate and high TIMP-2 expressing clones showed decreased growth as s.c. tumors [19]. The synthetic inhibitors SC-44463 [20], BEI6627B [21] and BB-94 [22] have all shown inhibition of experimental metastasis in mice. In the one study of spontaneous metastasis, BB-94 decreased the size of the lung metastases but had no effect on the number of metastases [22]. BB-94 administration also inhibited the growth of s.c. B16/F10 tumors [22], human colon tumors implanted in the mouse colon [23] and human ovarian tumors implanted intraperitoneally (i.p.) [24]. BE 16627B inhibited the growth of a s.c. tumor that had high MMP expression, with little effect on a tumor that had low MMP expression [21]. Taken together the data show that MMP inhibition can decrease lung colonization after i.v. injection of tumor cells, whereas effects on spontaneous metastasis appear minimal [22]. The variable effects of MMP inhibition on tumor establishment and growth, ranging from no effect [16] to pronounced inhibition [21] of growth, may reflect differences in the experimental tumors and the degree of MMP inhibition in vivo. Further development of MMP inhibitors as clinical candidates would be facilitated by a better understanding of the relationships between inhibitor potency against MMPs in vitro, plasma concentrations in vivo and efficacy against tumor growth and spontaneous metastasis. To investigate these relation116

Clinical & Experimental Metastasis Vol 14 No 2

ships four potent synthetic MMP inhibitors were delivered by s.c. minipump, achieving steady-state plasma inhibitor concentrations far in excess of those needed to inhibit purified MMP enzymes. Inhibitor delivery throughout the sequence of tumor growth and spontaneous metastasis to the lung slowed the growth of some of the tumors tested. Because of the possible importance of cathepsins in matrix degradation, the inhibitor E-64 (L-trans-epoxysuccinyl-L-leucylamido (4-guanidino)butane was also tested [4].

Materials and methods Cell lines and culture conditions Tissue culture reagents were from Gibco BRL (Gaithersburg, MD, USA), except nonessential amino acids (Irvine Scientific, Santa Ana, CA, USA). Mat Ly Lu rat prostate tumor (from Dr James L. Mohler, University of North Carolina, Chapel Hill, NC, USA) and M27 murine Lewis lung variant (National Cancer Institute, Frederick, MD, USA) were grown in RPMI 1640 media containing 10% FBS, 2 mM L-glutamine and 0.05 mg/ml gentamicin. HT-29 human colon adenocarcinoma (American Tissue Culture Collection, Rockville, MD, USA) was grown in McCoy's 5a media containing 10% FBS and 0.05 mg/ml gentamicin. PC-3 human prostate adenocarcinoma (American Tissue Culture Collection) was grown in DMEM/F12 media containing 10% FBS and 200 mM L-glutamine. B 16/F 10 murine melanoma (National Cancer Institute) was grown in DMEM high glucose media containing 10% FBS, 400 mM L-glutamine, 1 mM sodium pyruvate and nonessential amino acids. Cells were harvested with I mM EDTA/0.05% trypsin, washed in PBS and injected at 1 x 106 (Mat Ly Lu, HT-29, B16/FI0) or 1.5 x 10 6 (M27, PC-3) cells per mouse s.c. in the axillary region of the flank. LOX human melanoma (Southern Research Institute, Birmingham, AL, USA) was maintained as subcutaneous tumors by transfer of 3 mm 3 fragments into the axillary region of the flank. M M P assays

Recombinant human mature collagenase (MMP-1) and stromelysin (MMP-3) were expressed in E. coli, purified from inclusion bodies and refolded for 1 h prior to use. The human 92 kDa gelatinase (MMP-9) was purified as described previously [25]. Rat collagenase was a gift from Dr John J. Jeffrey (Department of Medicine, Albany Medical College, Albany, NY, USA). Collagenase assays were conducted in a buffer (200 mM NaC1, 50 mM Tris, 5 mM CaCIz, 20/~M ZnSO4, pH 7.6) containing 0.05% Brij 35, 3 nM enzyme and 100 #M of the chromogenic substrate

M M P inhibitors and tumor growth and metastasis Ac-Pro-Leu-GIy-SCH[CH2CH(CH3)2]CO-Leu-GlyOC2H 5 (Bachem Inc., Torrance, CA, USA) [26]. Human 92 kDa gelatinase was measured in assay buffer containing 0.01% Brij 35, 1 nM enzyme and 10 #M of the fluorogenic substrate Dnp-Pro-Cha-GlyCys(Me)-His-Ala-Lys(Nma)-NH2 [25]. Human stromelysin was measured in assay buffer containing 0.05% Brij 35, 10 nM enzyme and 10 /AM of the fluorogenic substrate Dnp-Pro-Gln-Gln-Phe-LysArg-Lys(Nma)-NH2. In all the above assays substrate concentrations are below the Km so the IC50 ~ Ki for the compounds. Cell proliferation assay Mat Ly Lu cells in RPMI 1640 media containing 10% FBS and 0.05 mg/ml gentamicin were added to 96-well fiat-bottom plates at 750 cells/100/Al/well. After 4 h 100 #1 of media containing 200/AM drug and 1% D M S O was added. Three days later the drug was removed, fresh media added for 3 h, and cell number assayed by the colorimetric tetrazolium assay [27]. Protease inhibitors and minipump implantation The M M P inhibitors are soluble analogs of a previously described series of potent collagenase [28] and gelatinase inhibitors [29]. The inhibitors were synthesized in the Department of Medicinal Chemistry at Glaxo Inc. E-64 was obtained from Sigma. For minipump delivery, all compounds except IV (50 mg/ml) were solubilized at 25 mg/ml. Compound I was dissolved in D M S O and added to 50 mM sodium citrate buffer, pH 4.0, giving a final DMSO concentration of 20% v/v. Compounds II, III and IV were dissolved in D M S O and added to vehicle for a final solution of 0.9% NaCI, 10%o D M S O and 31%o Tappsol HPB-20 (CTD Inc., Gainesville, FL, USA). E-64 was dissolved in 50% ethanol. Mice were anesthetized with an i.p. injection of 0.06 mg xylazine and 1.8 mg ketamine and two 7- or 14-day osmotic minipumps (Alza Corp., Palo Alto, CA, USA) containing drug or vehicle solutions implanted s.c. in the upper dorsum. The 7-day pump delivers 1 /A1/h whereas the 14-day pump delivers 0.5 /Al/h. For measurements of drug levels mice were euthanized with CO2, blood was drawn from the inferior vena cava with a heparinized syringe and plasma separated by centrifugation at 2000 g for 15 rain. Samples were frozen at - 20°C until assayed. Plasma assays of protease inhibitors The general approach was to remove plasma proteins and serial dilute the remaining fraction to get 50% inhibition of the target enzyme activity. The number of dilutions necessary to get 50% inhibition was

multiplied by the known IC50 concentration to calculate initial plasma drug concentrations. Compound I was separated from plasma by centrifugation through a 30000 MW cut-off filter (Model YMT-30; Amicon, Beverly, MA, USA) for 15 min at 1500 g. The filtrate was tested for collagenase inhibition (see M M P assays above). Control experiments showed no binding of the compound to plasma proteins or the filter apparatus. For measurement of compounds II, III and IV, 200 #1 of plasma was mixed with 100 /AI of 5% trichloroacetic acid, then centrifuged and 180 #1 of the supernatant neutralized by 5-10 #1 of 1 M Tris base. The supernatant was then tested for its ability to inhibit activated human 92 kDa gelatinase (see M M P assays above). The concentrations of compounds II, III and IV in the supernatant were multiplied by 4.8, 1.5 and 1.3, respectively, to correct for drug recovery through the plasma acidification-neutralization procedure. To measure E-64 plasma levels, plasma was centrifuged through a 30 000 MW cut-off filter for 15 rain at 1500 g. E-64 inhibition of papain activity was measured by modification of the procedures of Barrett et al. [30]. Plasma filtrate (50/A1) was mixed with 50/AI of 25 nM activated papain and 0.5 ml of a buffer of 0.017 r~ citrate/0.066 M dibasic sodium phosphate, pH 6.2. After a 5 min incubation, 0.4 ml of 20/AM Z-Phe-ArgAMC (Bachem Biosciences Inc., Philadelphia, PA, USA) substrate in buffer was added and the reaction terminated 10 min later with 1 ml of 100 mM sodium monochloroacetate in 150 mM sodium chloroacetate, pH 4.3. Product was measured fluorometrically using excitation and emission wavelengths of 370 and 460 riM, respectively. Serial dilutions of plasma filtrate alone were run to correct for inhibition by filtrate alone; undiluted filtrate gave about 30% inhibition. In control experiments E-64 did not bind to plasma proteins or the filter appparatus. Plasma binding of M M P inhibitors Inhibitors were added to plasma at 1/AMand allowed to sit for 20 min at room temperature. The plasma was then centrifuged through a 30 000 MW cut-off filter (Model YMT-30; Amicon) for 3 min at 1500 g. Less than 20% of the plasma appeared as filtrate. The concentration of inhibitor in the filtrate was then quantitated using coUagenase inhibition (see M M P assays above). Tumor growth and metastasis Primary s.c. tumors were measured with calipers three times weekly, and tumor volume calculated using the formula (length × width2)/2. After specific times of primary tumor growth the lungs were removed and Clinical & Experimental Metastasis Vol 14 No 2

117

C. Clemente et al. bioassayed for spontaneous metastasis by a modification of the method described by Dykes et al. [31]. Primary tumor-bearing mice were sacrificed by cervical dislocation, the lungs removed and rinsed in PBS containing 0.2% penicillin/streptomycin. The whole lung and a minimial amount of PBS was transferred to a 3 ml glass syringe with a 18-gauge needle, the air bubbles cleared and the lung injected s.c. into an untreated recipient mouse. The growth of secondary tumors from the implanted lung was measured with calipers three times weekly.

k. o

I

,-,

o

rF'o

HO.

I-(GI168) o

O

o

H

CA?

Results Determination of inhibitor potency in vitro The chemical structures of M M P inhibitors I, II, III and IV are shown in Figure 1. Forward (I, II, IV) and reverse (III) hydroxamic acid functions provide binding to the zinc in the enzyme active site. Compound I is an example of a P]-P~ leucine-phenylalanine substrate analogue of collagenase 1-28]. Compounds II, III and IV are examples of chemical series optimized for gelatinase inhibition [29]. Inhibition of human collagenase, gelatinase and stromelysin and rat collagenase is shown in Table 1. Compounds II, III and IV were the most potent against human gelatinase and showed selectivity relative to human collagenase. To further explore gelatinase inhibition across species and tumor types, zymograms of serum-free media from human LOX melanoma, rat Mat Ly Lu prostate tumor and mouse M27 Lewis lung carcinoma ceils were used to demonstrate the expression of gelatinases (data not shown). When the zymograms were renatured and developed in the presence of 50 nM compound I or 10 nM compounds II, III and IV, gelatinases from all three cell lines were inactive (data not shown), indicating that at these concentrations complete inhibition of each species of gelatinase could be achieved. Measurement of plasma drug concentrations after minipump implantation Infusion of the inhibitors into mice with s.c. minipumps resulted in steady-state plasma concentrations well above the concentrations needed for inhibition of specific M M P in vitro (compare Tables 1 and 2). For example, infusion of compound I in 7-day minipumps resulted in plasma concentrations 158-, 158- and 2.2-fold higher than the concentrations needed for 50% inhibition of human collagenase, gelatinase B and stromelysin, respectively. Thus, with this relatively weak stromelysin inhibitor one would expect minimal inhibition of stromelysin in vivo. However, compound II shows different selectivity towards MMPs: infusion 118

Clinical & Experimental Metastasis Vol 14 No 2

II-(GI173) OH H

O

-

H

O

III-(GI179) 0 ~

o

H r/"N"~

°, 0 IV-(GI184) Figure 1. Chemical structures of the matrix metalloproteinase inhibitors (compounds I-IV).

Table 1. Potencies of inhibitors I-IV a against matrix

metalloproteinase in vitro Drug

Inhibition IC5o (nM) Human Collagenase

I II III IV

3 37 149 26

Rat collagenase

Gelatinase Stromelysin 3 0.05 0.1 0.06

a Structures shown in Figure 1.

225 8 14 16

15 3 19 4

M M P inhibitors and tumor growth and metastasis in 7-day minipumps resulted in plasma concentrations 16-, 11 600- and 73-fold higher than the concentrations needed for 50% inhibition of human collagenase, gelatinase B and stromelysin, respectively. Extremely high binding to plasma proteins coupled with a slow off-rate could limit the distribution of inhibitors into tissue sites of M M P activity. The

Table 2. Drug plasma concentrations in normal mice with two minipumps implanted subcutaneously

Drug a

Pump type

Drug delivery (mg/kg/day) b

Day of sample

Plasma conc. (nM)c

n

I

7-day

48

II

14-day 7-day

24 48

III

14-day

24

IV

14-day

48

7-day 7-day

96 48

2 7 2 2 7 7 14 7 14 7 7

565 _+72 390_+96 119_+18 491 _+89 665 _+272 125 _+16 81_+16 362_+ 111 262_+67 794 + 94 5190_+1006

5 5 6 5 4 7 7 4 4 5 5

degrees of plasma binding for compounds I, II, III and IV were 10, 73, 62 and >95% , respectively. Thus, it seems unlikely that plasma binding would limit the distribution of compounds I, II and III. Given that compound IV shows anti-tumor activity in vivo (next section) it is reasonable to assume its protein binding is reversible. E-64, an irreversible inhibitor of cysteine proteases, was maintained in the plasma at 5.2/ZM by means of 7-day minipumps (Table 2). Kinetic studies with isolated enzymes show that cathepsins B, H and L are inhibited within seconds with this concentration of inhibitor [30, 32].

Bioassay of spontaneous metastasis to the lung

E-64

a Structures for compounds I, II, III and IV shown in Figure 1. E-64=irreversible inhibitor L-trans-epoxysuccinyl-Lleucylamido(4-guanidino)butane. b Calculated from flow rates of 1 and 0.5/A for 7- and 14-day pumps, respectively. All pumps contained 25 mg/ml drug, except IV which was at 50 mg/ml. cData represent mean _+SEM ofn mice. Drug concentrations measured by inhibition of collagenase activity.

LOX, Mat Ly Lu and M27 tumors growing s.c. spontaneously metastasized to the lung in 10-12 days. Metastasis was bioassayed by injecting the lungs of the tumor-bearing mice s.c. into untreated mice, and measuring the outgrowth of the secondary tumor from the implanted lung. Time-course experiments with these three tumors are shown in Table 3. Primary tumors grew rapidly over days 2-12. At different times during the primary tumor growth the lungs were assayed for metastasis. With all three tumors the incidence and rate of outgrowth of secondary tumors from the implanted lungs increased as the primary tumor grew out from days 2-3 to days 10-12 (Table 3), indicating that the metastasic burden of the lung increases as the primary tumor grows over this period of time. Thus, a significant inhibition of metastasis should decrease the incidence and outgrowth of the secondary tumors.

Table 3. Relationship between day of primary tumor growth and outgrowth of secondary tumors from bioassay lungs

Tumor typea

Mat Ly Lu

LOX

M27

Day of lung transfer

2 5 10 3 7 12 2 6 10

Primary tumor size (mm3)b 50_+ 12 204-t-39 1437___147 45 _+17 437+45 2001 + 213 43 -t-4 140_+ 15 702 _+39

Secondary tumors ¢ Incidence

Time to reach 1 g (days)d

1/10 4/10 10/10 2/10 6/10 10/11 1/10 10/10 10/10

30 18+2 10__+1 27 20_+ 1 14 _+1 26 21 _ 1 14_+ 1

a Mat Ly Lu rat prostate tumor, LOX human melanoma and M27 murine Lewis lung tumor. b Data represent mean_ SEM of 10 primary tumors at the day of lung transfer s.c. into an untreated mouse. Tumors growing out of the lungs implanted s.c. into untreated mice. d Data represent mean + SEM of the secondary tumors that grew out.

Clinical & Experimental Metastasis Vol 14 No 2

119

C. Clemente et al. Effect o f druo treatments on tumor 9rowth and spontaneous metastasis Minipump delivery over 7 or 14 days was used to test the effects o f M M P inhibitors on the growth of various s.c. tumors in mice. The 7-day pumps were implanted at the beginning of rapid tumor growth and the 14-day pumps were implanted 1-2 days before tumor implantation. Two typical experiments showing inhibition of the growth of May Ly Lu tumors in mice receiving 48 mg/kg/day of compounds I and II from days 3-10 are shown in Figure 2. In all six experiments

with the Mat Ly Lu rat prostate tumor M M P inhibitors caused statistically significant inhibition of primary tumor growth (Table 4). The degree of inhibition b y compound I in 7-day (48 mg/kg/day) or 14-day (24 mg/kg/day) minipumps was similiar at 40-60%. The other three inhibitors gave similiar results with this tumor. The very slow growing human prostate adenocarcinoma PC-3 was also inhibited by 44% in the one experiment conducted (Table 4). Growth of human LOX melanoma tumors was inhibited by 23, 44, 57 and 61% in four separate

Table 4. Effect of matrix metalloproteinase inhibitors on primary tumor growth and spontaneous metastasis to the lung

Tumor type/ druga

Mat Ly Lu I I II II III IV E-64 LOX I I III IV M27 I II IV IV E-64 E-64 HT29 I PC-3 I BI6/FI0 Ie

Dose (mg/kg/day)

Treatment days

Primary tumors % Inhibition of growthb

Size at lung transfer ~

Secondary lung tumors Incidence d

Time taken to reach 1 g (days)

C

T

C

T

24 48 48 48 24 48 24

-2-12 0-8 3-10 3-10 -2-12 - 1-12 -1-12

43* 49*** 57*** 57** 43** 38** 11

1470+231 568___54 1242+__180 1523_+132 1681_+145 1636-+ 135 1790_+173

7/7 10/10 4/7 9/9 6/7 6/6 6/6

10/10 9/10 6/8 9/10 10/10 10/10 10/10

11 _+1 19_+2 15+_3 12_+1 13_+1 7-+ 1 7_+0

12_+0 16+1 16_+ ! 11_+1 13-+1 9-+ 1 8_+1

48 24 24 96

5-12 -2-11 - 1-13 3-10

23 61"** 44 58

2378+384 4152-+271 1440_+401 705 + 173

6/7 8/10 8/9 8/9

10/10 6/10 5/8 7/10

17-t-2 15_+1 25_+ 1 22 +__2

18-t-1 18_+1 21 _+4 32_+2**

48 48 48 96 48 24

5-12 3-10 -2-12 2-9 5-12 -2-12

21 30 31"** 10 0 0

305 _+27 699_ 74 521+28 215_+23 296 _+13 387-+35

10/10 8/10

10/10 10/10

18_+2 17+__2

16+1 19_+2

8/9

9/9

19-+2

16+1

48

17-24

13

397 _ 24

24

10-25

44**

74-+ 8

48

6-13

0

668 _+102

aTumors were Mat Ly Lu rat prostate tumor, LOX human melanoma, M27 murine Lewis lung tumor, HT-29 human colon adenocarcinoma, human prostate adenocarcinoma PC-3 and B16/F10 murine melanoma. Compound structures shown in Figure 1. E-64--irreversible inhibitor L-trans-epoxysuccinyl-L-leucylamido(4-guanidino)butane. b Measured at day of lung transfer. c Data represent mean_+ SEM of tumors in vehicle-treated mice. d Positive lung metastasis/total mice in the group; C = control; T = treated. e Performed using C57/BL mice. *P < 0.05; **P < 0.01; ***P < 0.001 compared with vehicle-treated controls using 2-sided t-test.

120

Clinical & Experimental Metastasis Vol 14 No 2

M M P inhibitors and tumor growth and metastasis (a) 160o 1400 E

E 6

12oo 1000

E -o >

800

600

E I--

400 200

0 (b) 1800

I

I

I

I

1600

E

1400

E

1200

E

1000

o >

800

6

inhibited the growth of Mat Ly Lu tumors by 57% (Figure 2). Drug treatment showed no obvious effect on the number of tumor blood vessels, but did increase the percentage of that tumor that was necrotic, even though the drug-treated tumors were smaller. M M P inhibitors administered over the time-frame of metastasis using either 7- or 14-day minipumps had no effect on the incidence of secondary tumors, even in experiments with Mat Ly Lu and LOX tumors where the drug treatment inhibited primary tumor growth by 50% (Table 4). In only one of 14 metastasis experiments did compound treatment inhibit the rate of outgrowth of the secondary tumors. In this experiment, using a LOX tumor, compound IV delivered at the highest dose (96 mg/kg/day) and maintained at the highest plasma concentration (Table 2) caused a statistically significant decrease in the rate of secondary tumor outgrowth (Table 4), indicating that some inhibition of metastasis had occurred.

m

E I.-

Discussion

600 400 200

4

5

6

7

8

9

10

Days of T u m o r Growth

Figure 2. Effect of (a) compound I and (b) compound II, given at 48 mg/kg/day on days 3-10, on the growth of Mat Ly Lu rat prostate tumors. In each case the compound ([-]) is compared with vehicle (0). See rows 3 and 4 of Table 4 for corresponding metastasis data. experiments. However, in only one of the four experiments was the inhibition statistically significant. M M P inhibitors had no significant effect on the growth of the murine Lewis lung variant M27, human colon tumor HT-29 or murine melanoma B16/F10 tumors. The cysteine protease inhibitor E-64 was tested against two lines (Mat Ly Lu and M27) and showed no significant effect on tumor growth (Table 4). None of the drug treatments showed any obvious toxicity or effect on body weight. Because of the observed inhibition of tumor growth, M M P inhibitors were tested in cell proliferation assays in vitro. Compound I given at 100 #M at the beginning of the 3-day assay had no effect on the proliferation of Mat Ly Lu, M27 or LOX cells (data not shown). Histological analysis was performed in two studies, where delivery of compounds I and II from days 3-10

Extracellular matrix degradation is a local phenomenon where degradative enzymes such as the M M P s are found in the local interstitial fluid and associated with the surfaces of cells. Both tumor and local stromal cells have been shown to produce M M P at sites of matrix degradation [6-11, 33]. Thus, one factor in assessing the efficacy of synthetic M M P inhibitors in vivo is the penetration of the drug to the local site. To try and maximize inhibitor concentrations at the site of matrix degradation, the small molecular weight inhibitors (MW 450-600) were constantly infused by s.c. minipumps. One would expect these small inhibitors to distribute throughout the body during constant 1- and 2-week infusions. Comparison of steady-state plasma drug concentrations (Table 2) and concentrations needed to inhibit purified collagenase, gelatinase and stromelysin in vitro (Table 1) shows that the inhibitor concentrations in vivo are far in excess of those required to inhibit the enzyme(s). Given possible differences in species and tumor types, tumors of different species (human, rat and mouse) and tumor histotypes (prostate, melanoma, lung adenocarcinoma, colorectal) were utilized to assess the efficacy of these M M P inhibitors. Drug treatment resulted in partial inhibition of the growth of the Mat Ly Lu rat prostate tumors and LOX melanoma tumors, as well as inhibition of the growth of the human PC-3 prostate tumor in the one experiment conducted (Table 4). Along with recent data that compound I inhibits adjuvant arthritis in rats Clinical & Experimental Metastasis Vol 14 No 2

121

C. C l e m e n t e et al.

at 150-250 nM in the plasma [34], this inhibition of tumor growth would suggest the inhibitor plasma levels achieved after minipump infusion (Table 2) are sufficient to inhibit M M P s in vivo. The inhibition of tumor growth in vivo is unlikely to be due to direct inhibition of cell proliferation since 100 ~tM of compound I had no effect on cell proliferation in vitro (data not shown), even though it inhibited tumor growth in vivo at plasma concentrations of 500 nM or less (Tables 2 and 4). BB-94, a potent M M P inhibitor structurally similiar to compound I, also had no effect on the proliferation of numerous cell lines [24]. Whereas the mechanism of tumor growth inhibition does not involve direct inhibition of cell proliferation, it is still possible that inhibition of matrix degradation could inhibit tumor angiogenesis and thus tumor growth. This mechanism is consistent with histological analysis suggesting a greater fraction of necrotic tissue in drug treated vs control tumors (see above). This mechanism is also supported by the observations that GM6001 inhibited corneal angiogenesis induced by tumor extract [35] and that BB-94 inhibited angiogenesis and the growth of hemangiomas in vivo [36]. Comparison of the inhibitor concentrations in plasma, the in vitro ICsos and the inhibition of tumor growth allows one to speculate as to which M M P was most important in the tumor growth inhibition. Compound I, a weak inhibitor of stromelysin (IC50 = 225 nM), compared with collagenase (IC5o = 3 nM) and gelatinase B (IC50=3 nM), inhibited the growth of Mat Ly Lu and LOX tumors when infused in 14-day pumps (Table 4). Inhibitor plasma concentrations with the 14-day pumps were only about 120 nM (Table 2), thus arguing against stromelysin inhibition being the most important in the inhibition of Mat Ly Lu and LOX tumor growth with compound I. In contrast, compound III is a weak inhibitor of collagenase (IC50 = 149 riM) compared with gelatinase B (IC50=0.1 nM) and stromelysin (IC50=16 nM). Compound III in 14-day minipumps inhibited the growth of May Ly Lu and LOX tumors at plasma concentrations of only about 100 nM (Tables 2 and 4). This result argues against collagenase being important in the inhibition of tumor growth with compound III. All four inhibitors are potent against gelatinase B (ICsos = 3 riM), and in all tumor growth experiments the plasma concentrations of inhibitor were many fold above the IC50 for gelatinase inhibition in vitro (Table 1 vs Table 2). Thus, unlike the possible incomplete inhibition of stromelysin with compound I and collagenase with compound III in vivo, gelatinase inhibition was universally associated with the inhibition of the growth of the Mat Ly Lu and LOX 122

Clinical & Experimental Metastasis Vol 14 No 2

tumors, and may be the most important target. Even so, we judge the anti-tumor effect to be rather modest given the high plasma levels and the potencies of the inhibitors used. This suggests that the M M P s studied here may not be essential for tumor growth. In most models of spontaneous metastasis the primary tumor must be grown for a long period and reach a large size before spontaneous lung metastasis can be quantified macroscopically. To avoid a long period of tumor growth, primary tumors have been resected, allowing the early lung metastasis to grow and permitting macroscopic quantification [22]. However resection of primary tumors can be difficult and causes significant trauma to the animal. An advantage of the bioassay of spontaneous metastasis is that metastasis can be detected at early timepoints, thus limiting the period of primary tumor growth and drug treatment. The ability to detect early metastasis is probably facilitated by the presence of co-injected lung tissue, as shown in previous studies where co-injection of nonreplicating tumor [37, 38] and lung [39] dramatically decreased the number of tumor cells required to initiate a tumor. The bioassay of lung metastasis was initially described by Dykes et al. [31]. Like us, they observed that the incidence and rate of outgrowth of secondary tumors from implanted lungs correlated with time of growth of the primary tumor. In addition, they showed that the treatment of the tumor-bearing mice with the cytotoxins, cyclophosphamide or dimethyltriazeno-imidazolecarboxyamide, just before lung transfer caused a dosedependent decrease in the incidence and rate of outgrowth of secondary tumors from the lungs implanted s.c. Taken together, the data show that incidence and rate of outgrowth of secondary tumors are indicative of the metastatic burden of the lung on the day of lung transfer to the recipient bioassay animal. In our study inhibitor treatment had no effect on spontaneous metastasis even with the Mat Ly Lu and LOX tumors where inhibitor treatment caused a significant decrease in primary tumor growth. The absence of a consistent effect of M M P inhibitors on spontaneous metastasis in our model system argues against an obligatory role for M M P s in this model. Similarly, Chirivi et al. [22] reported that B-94 did not decrease the number of spontaneous metastases to the lung. These results in spontaneous metastasis models are in contrast to the inhibition of experimental metastasis observed with biological [17-19] and synthetic M M P inhibitors [20-22]. It is possible that our bioassay read-out of spontaneous metastasis might require a more significant decrease in lung metastatic tumor burden to show statistically

M M P inhibitors and tumor growth and metastasis

significant decreases in secondary tumor incidence and rate of outgrowth. It is also possible that the subset of tumor cells that can successfully metastasize in the spontaneous model are not dependent on M M P s and may use other mechanisms. E-64 at the plasma concentrations achieved by minipump infusion (Table 2) will rapidly cause irreversible inhibition of cysteine proteases cathepsin B, H and L [30, 32]. This compound has a limited ability to enter cells [40], so in vivo it will preferentially inhibit the extracellular cysteine proteases. The observation that E-64 had no effect on tumor growth or spontaneous metastasis of Mat Ly Lu and M27 tumors suggests that extracellular cathepsins B, H and L play a minor role in these model systems. These data are also consistent with the report by Ostrowski et al. [41] that minipump infusion of the prodrug ester of E-64, E-453, had no effect on the ability of B16/F10 melanoma cells to lodge and grow in the lung after i.v. injection.

Acknowledgement We would like to thank Mr Neil Jones for conducting the assays of cell proliferation.

References 1. Folkman J and Shing Y, 1992, Angiogenesis. J Biol Chem, 16, 10931-4. 2. Liotta LA, 1986, Tumor invasion and metastasis: role of the extracellular matrix. Cancer Res, 46, 1-7. 3. Ray JM and Stetler-Stevenson WG, 1994, The role of matrix metalloproteinases and their inhibitors in tumour invasion, metastasis and angiogenesis. Eur Respir J, 7, 2062-72. 4. Sloane BF, Moin K, Krepela E and Rozhin J, 1990, Cathepsin B and its endogenous inhibitors: their role in tumor malignancy. Cancer Metastasis Rev, 9, 333-52. 5. Wolf C, Chenard M, Durand de Grossouvre P, Bellocq J, Chambon P and Basset P, 1992, Breast cancerassociated stromelysin-3 gene is expressed in basal cell carcinoma and during cutaneous wound healing. J Invest Dermatol, 99, 870-2. 6. Pyke C, Ralfkiaer E, Huhtala P, Hurskainen T, Dano K and Tryggvason K, 1992, Localization of messenger RNA for Mr 72,000 and 92,000 type IV collagenases in human skin cancers by in situ hybridzation. Cancer Res, 52, 1336-41. 7. Polette M, Clavel C, Cockett M, Girod De Bentzmann S, Murphy G and Birembaut P, 1993, Detection and localization of matrix metalloproteinases and their tissue inhibitor in human breast pathology. Invasion Metastasis, 13, 31-7. 8. Stearns ME and Wang M, 1993, Type IV collagenase (Mr 72,000) expression in human prostate: benign and malignant tissue. Cancer Res, 53, 878-83.

9. Wolf C, Rouyer N, Lutz Y, et al. 1993, Stromelysin 3 belongs to a subgroup of proteinases expressed in breast carcinomal fibroblastic cells and possibly implicated in tumor progression. Proc Natl Acad Sci USA, 90, 1843-7. 10. Kawame H, Toshida K, Ohsaki A, Kuroi K, Nishiyama M and Tage, T, 1993, Stromelysin-3 mRNA expression and malignancy: comparison with clinicopathogical features and type IV collagenase mRNA expression in breast tumors. Anticancer Res, 13, 2319-24. 11. Garbisa S, Pozzatti R, Muschel RJ, et al. 1987, Secretion of type IV collagenolytic protease and metastatic phenotype: induction of transfection with c-Ha-ras but not c-Ha-ras plus Ad2-Ela. Cancer Res, 47, 1523-8. 12. Sreenath T, Matrisian LM, Stetler-Stevenson W, Gattoni-Celli S and Pozzatti RO, 1992, Expression of matrix metalloproteinase genes in transformed rat cell lines of high and low metastatic potential. Cancer Res, 52, 4942-7. 13. Bernhard EJ, Muschel RJ and Hughes EN, 1990, Mr 92,000 Gelatinase releases correlates with the metastatic phenotype in transformed rat embryo cells. Cancer Res, 50, 3872-7. 14. Terranova VP, Hujanene ES, Loeb DM, Martin GR, Thornburg L and Glushko V, 1986, Use of a reconstituted basement membrane to measure cell invasiveness and select for highly invasive tumor cells. Proc Natl Acad Sci USA, 83, 465-9. 15. Nakajima M, Welch DR, Belloni PN and Nicolson GL, 1987, Degradation of basement membrane type IV collagen and lung subendothelial matrix by rat mammary adenocarcinoma cell clones of differing metastatic potentials. Cancer Res, 47, 4869-76. 16. Schultz RM, Silberman S, Persky B, Bajowski AS and Carmichael DF, 1988, Inhibition by human recombinant tissue inhibitor of metalloproteinases of human amnion invasion and lung colonization by murine B16-F10 melanoma cells. Cancer Res, 48, 5539-45. 17. Alvarez OA, Carmichael DF and DeClerck YA, 1990, Inhibition of collagenolytic activity and metastasis of tumor cells by a recombinant human tissue inhibitor of metalloproteinases. J Natl Cancer Inst, 82, 589-95. 18. Khokha R, 1994, Suppression of the tumorigenic and metastatic abilities of m urine B 16-F 10 melanoma cells in vivo by the overexpression of the tissue inhibitor of the metalioproteinases-1. J Natl Cancer Inst, 86, 299-304. 19. DeClerck YA, Perez N, Shimada H, Boone TC, Langley KE and Taylor SM, 1992, Inhibition of invasion and metastasis in cells transfected with an inhibitor of metalloproteinases. Cancer Res, 52, 701-8. 20. Reich R, Thompson EW, Iwanoto Y, et al. 1988, Effects of inhibitors of plasminogen activator, serine proteinases, and collagenase IV on the invasion of basement membranes by metastatic cells. Cancer Res, 48, 3307-12. 21. Naito K, Kanbayashi N, Nakajima S, et al. 1994, Inhibition of growth of human tumor cells in nude mice by a metalloproteinase inhibitor. Int J Cancer, 58, 730-5. 22. Chirivi RGS, Garofalo A, Crimmins M J, et al. 1994, Inhibition of the metastatic spread and growth of B16-BL6 murine melanoma by a synthetic matrix metalloproteinase inhibitor. Int J Cancer, 58, 460-4. 23. Wang X, Fu X, Brown PD, Crimmin MJ and Hoffman RM, 1994, Matrix metalloproteinase inhibitor BB-94 (Batimastat) inhibits human colon tumor growth and

Clinical & Experimental Metastasis Vol 14 No 2

123

C. Clemente et al.

24.

25.

26. 27. 28. 29.

30.

31.

32.

124

spread in a patient-like orthotopic model in nude mice. Cancer Res, 54, 4726-8. Davies B, Brown PD, East N, Crimmin MJ and Balkwill FR, 1993, A synthetic matrix metalloproteinase inhibitor decreases tumor burden and prolongs survival of mice bearing human ovarian carcinoma xenografts. Cancer Res, 53, 2087-91. Bickett DM, Green MD, Berman J, et al. 1993, A high throughput fluorogenic substrate for interstitial collagenase (MMP-1) and gelatinase (MMP-9). Anal Biochem, 212, 58-64. Weingarten H, Martin R and Feder J, 1985, Synthetic substrates of vertebrate collagenases. Biochemistry, 24, 67304. Mosmann T, 1983, Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J Immunol Methods, 65, 55-63. Campion C, Dickens JP and Crimmin M J, 1990, Patent WO 90/05719, British Biotechnology. Porter JR, Morphy JR, Millican TA and Beeley NRA, 1993, New sulphamide derivatives are metalloproteinase inhibitors for inhibiting tumor metastasis. Patent WO 9324475-A1, Cell Tech Ltd. Barrett A J, Kembhavi AA, Brown MA, et al. 1982, L-trans-Epoxysuccinyl-leucylamido(4-guanidino)butane (E-64) and its analogues as inhibitors of cysteine proteinases including cathepsins B, H, and L. Biochem J, 201, 189-98. Dykes D, Shoemaker R, Harrison S, et al. 1987, Development and therapeutic response of a spontaneous metastasis model of a human melanoma. Proc Am Assoc Cancer Res, 28, Abstract 1709. Hashida S, Towatari T, Kominami E and Katunuma N, 1980, Inhibitions by E-64 derivatives of rat liver cathepsin B and cathepsin L in vitro and in vivo. J Biochem, 88, 1805-11.

Clinical & Experimental Metastasis Vol 14 N o 2

33. Miyagi E, Yasumitsu H, Hirahara F, et al. 1995, Marked induction of gelatinases, especially type B, in host fibroblasts by human ovarian cancer cells in athymic mice. Clin Exp Metastasis, 13, 89-96. 34. Conway JG, Wakefield JA, Brown RH, et al. 1995, Inhibition of cartilage and bone destruction in adjuvant arthritis in the rat by a matrix metalloproteinase inhibitor. J Exp Med, 182, 449-57. 35. Galardy RE, Grobelny D, Foellmer HG and Fernandez LA, 1994, Inhibition of angiogenesis by the matrix metalloproteinase inhibitor N-[2R-2-(hydroxamidocarbonymethyl) - 4 - methylpentanoyl)] - L - tryptophan methylamide. Cancer Res, 54, 4715-18. 36. Taraboletti G, Garafalo A, Belotti D, et al. 1995, Inhibition of angiogenesis and murine hemagioma growth by batimastat, a synthetic inhibitor of matrix metalloproteinases. J Natl Cancer Inst, 87, 293-8. 37. Dykes DJ, Griswold DP, Jr and Schabel FM, 1976, Growth support of small B16 melanoma implants with nitrourea-sterilized fractions of the same tumor. Cancer Res, 36, 2031-4. 38. Hewitt HB, Blake E and Porter EH, 1973, The effect of lethally irradiated cells on the transplantability of murine tumours. Br J Cancer, 28, 123-35. 39. DeWys WD, 1972, A quantitative model for the study of the growth and treatment of a tumor and its metastasis with correlation between proliferative state and sensitivity to cyclophosphamide. Cancer Res, 32, 367-73. 40. Wilcox D and Mason R, 1992, Inhibition of cysteine proteinases in lysosomes and whole cells. Biochem J, 285, 495-502. 41. Ostrowski LE, Ahsan A, Suthar BP, et aL 1986, Selective inhibition of proteolytic enzymes in an in vivo mouse model for experimental mestastasis. Cancer Res, 46, 4121-8.